Early warning indicators can help manage credit risk
Banks use a number of approaches to manage credit risk. Despite this, banks continue to suffer from squeezing of profitability and spiraling cost of operations. One of the primary reasons for this could be use of manual or semi-automated approaches in credit risk monitoring, even in large banks.
The COVID-19 pandemic aggravated the problem across the globe. To manage the problem better, we suggest adoption of intelligent tools in risk operations – namely, early warning indicators (EWI) that leverage natural language processing and artificial intelligence techniques.
With the use of EWI, positive and negative connotations contained in unstructured form of data, such as real-time media reports and newsfeeds can be automatically converted into a risk rating or score for generating alerts.
How EWI works
The following are examples where EWI automatically interprets newsfeeds for generating alerts by using two parts of a May 2020 report from the U.S. Bureau of Economic Analysis.
- Gross domestic product for the first quarter of 2020, second revision, suggested a decrease in real GDP by at an annual rate of 5 percent. This compared to increase of 2.1 percent in the fourth quarter of 2019.
- Interpretation and alert generation: Since this is a drastic fall, this is highly negative from the view point of credit risk. This newsfeed would be automatically interpreted by EWI and an alert would be generated to key stakeholders
- A preliminary estimate of profits for domestic, nonfinancial corporations decreased $169.5 billion in the first quarter of 2020, in contrast to an increase of $53.7 billion in the final quarter of 2019.
- Interpretation and alert generation: Decreased profits is a sign of increase in credit risk. Given the size of fall, this would be automatically interpreted by EWI as highly negative and an alert would be generated to key stakeholders
Relevance of EWI in managing COVID-19 challenges
Since the coronavirus crisis began, many banks had difficulties or were even unable to conduct lending and credit risk management processes despite triggering business continuity plans.
In addition to delayed processing of pending loan requests, banks suffered the most when their manual or semi-automated credit risk monitoring and control activities, such as periodic review of health of borrowers, facilities provided to them and stress testing of collateral values came to a grinding halt. This led to further deterioration in credit quality, increase in credit losses and loan-loss provisions. This showed up as reduction in stock prices of banks that were publicly traded.
In those banks that had intelligent tools such as EWI, the position might have been much different. In the case of banks with mature EWI capabilities, an alert would have been automatically generated by EWI on the crisis much earlier than it could be manually interpreted and the alert generated. Even EWI with limited capabilities is better than the manual system of interpretation.
As a result, banks using EWI could have been positioned to make better decisions faster and prevented or limited credit losses. Even after normalcy returns, EWI is highly useful as a tool for early detection of potential drivers of losses and to initiate remedial actions.
Global experience in management of credit risk showed repeatedly that proactive monitoring is the difference between success and failure. Banks are increasingly adopting cognitive tools to replace manual approaches in credit risk management and especially in credit risk monitoring.
Upon attaining certain level of maturity, EWI could serve as a valuable source of early detection of risks in credit portfolios of banks. Moreover, this can serve as inputs for stress testing and recovery planning systems of banks.